A Case Study of the IBM SP2 Eric L

Total Page:16

File Type:pdf, Size:1020Kb

A Case Study of the IBM SP2 Eric L Modeling Computation and Communication Performance of Parallel Scientific Applications: A Case Study of the IBM SP2 Eric L. Boyd, Gheith A. Abandah, Hsien–Hsin Lee, and Edward S. Davidson Advanced Computer Architecture Laboratory Department of Electrical Engineering and Computer Science University of Michigan 1301 Beal Avenue Ann Arbor, MI 48109-2122 PHONE: 313-936-2917; FAX: 313-763-4617 {boyd, gabandah, linear, davidson}@eecs.umich.edu Abstract A methodology for performance analysis of Massively Parallel Processors (MPPs) is presented. The IBM SP2 and some key routines of a finite element method application (FEMC) are used as a case study. A hierarchy of lower bounds on run time is developed for the POWER2 processor, using the MACS methodology developed in earlier work for uniprocessors and vector processors. Significantly, this hierarchy is extended to incorporate the effects of the memory hierarchy of each SP2 node and communication across the High Performance Switch (HPS) linking the nodes of the SP2. The performance models developed via this methodology facilitate explaining performance, identifying performance bottlenecks, and guiding application code im- provements. 1. Introduction Scientific applications are typically dominated by loop code, floating–point operations, and array references. The perfor- mance of such applications on scalar and vector uniprocessors has been found to be well characterized by the MACS model, a hierarchical series of performance bounds. As depicted in Figure 1, the M bound models the peak floating–point performance of the Machine architecture independent of the application. The MA bound models the machine in conjunction with the opera- tions deemed to be essential in the high–level Application workload. Hence Gap A represents the performance degradation due to the application algorithm’s need for operations that are not entirely masked by the most efficient class of floating–point op- erations. The MAC bound is derived from the actual Compiler–generated workload. Hence Gap C represents the performance degradation due to the additional operations introduced by the compiler. The MACS bound factors in the compiler–generated Schedule for the workload. Gap S thus represents the performance degradation due to scheduling constraints. The performance degradation seen in the remaining gap, Gap P, is due to as yet unmodeled effects, such as unmodeled cache miss penalties, load imbalances, OS interrupts, and I/O, that affect actual delivered performance. The bounds are generally expressed as lower bounds on run time and, along with measured run time, are given in units of CPF (clocks per essential floating–point operation). The reciprocal of CPF times the clock rate in MHz yields an upper bound on performance in MFLOPS. The MACS model has been effectively demonstrated and applied in varying amounts of detail to a wide variety of processors. [1][2][3][4][5][6] The MACS model developed for each SP2 node incorporates a significant additional refinement beyond these earlier studies by in- cluding the effects of the memory hierarchy. Gap A Gap C Gap S Gap P Measured M MA MAC MACS CPF Figure 1: MACS Performance Bound Hierarchy Scientific applications often exhibit a large degree of potential parallelism in their algorithms, and hence are prime candi- dates for execution on MPPs. Gap P can become very large and poses the fundamental limit to scalability for parallel applica- tions. Characterizing the communication performance of an MPP and the communication requirements of an application achieves a refinement of Gap P that is crucial to modeling and improving the performance of parallel scientific applications that exhibit moderate to high amounts of communication relative to computation. We demonstrate that coupling the MACS bounds hierarchy, which models the computation of an individual node, with models of communication, as well as load balancing and cache effects, to refine Gap P, enables the effective modeling of the performance of scientific applications on a parallel com- puter such as the IBM SP2. This refinement is begun in this paper with the introduction of a communication model for the SP2. The MACS model is applied to the SP2 and integrated with this communication model in a case study of a commercial scientific 1 This document was created with FrameMaker 4.0.4 application. This methodology could easily be extended to other message passing MPPs, such as the Intel Paragon and the Thinking Machines CM5. [7][8] We believe that it can also be extended to shared memory MPPs, such as the Kendall Square Research KSR2, the Convex Exemplar, and the Cray T3D [11][12][13][14], by using techniques to expose and characterize the implicit communication, as demonstrated in [6][9][10]. All experiments in this paper were run on an IBM SP2 with 32 Thin Node 66 POWER2 processors running the AIX 3.2.5 operating system. An overview of the SP2 architecture is given in Section 2. The single–node SP2 (POWER2) MACS model is detailed in Section 3, with extensions to include the effects of the memory hierarchy. Section 4 presents the communication model for the IBM SP2. An industrial structural finite element modeling code, FEMC, is used to demonstrate the methodology. This application is being ported from a vector supercomputer, parallelized, and tuned for the IBM SP2 at the University of Michigan, and hence represents one scenario of performance modeling. Three performance–limiting FEMC routines are examined. These routines exhibit low, moderate, and high communication to computation ratios and various patterns of communication. Section 5 gives a brief overview of the application routines and discusses the performance modeling results. The performance models devel- oped via this methodology facilitate explaining performance, identifying performance bottlenecks, and guiding application code improvements. 2. IBM SP2 Architectural Analysis A typical IBM SP2 contains between 4 and 128 nodes connected by a High–Performance Switch communication intercon- nect; bigger configurations are possible. Each node consists of a POWER or POWER2 processor and an SP2 communication adapter. There are three types of POWER2 nodes currently available: Thin Nodes, Thin 2 Nodes, and Wide Nodes. Thin Nodes have a 64 Kbyte data cache; Thin 2 Nodes have a 128 Kbyte data cache, and Wide Nodes have a 256 Kbyte data cache. Thin 2 Nodes and Wide Nodes can do quadword data accesses, copying two adjacent double words into two adjacent floating point registers from the primary cache in one cycle. 2.1. POWER2 Architecture [15] Each Thin Node POWER2 operates at a clock speed of 66.7 MHz, corresponding to a 15 ns processor clock cycle time. The POWER2 processor is subdivided into an Instruction Cache Unit (ICU), a Data Cache Unit (DCU), a Fixed–Point Unit (FXU), and a Floating–Point Unit (FPU), as shown in Figure 2. The POWER2 processor also includes a Storage Control Unit (SCU) and two I/O Units (XIO), neither of which are shown or described. Thin Nodes and Thin 2 Nodes also include an optional sec- ondary cache (1 or 2 Mbytes, respectively), memory (64 Mbytes to 512 Mbytes), and a communication adapter. All experiments in this paper are run on Thin Nodes with no secondary cache and a 256 Mbyte memory. The ICU can fetch up to eight instructions per clock cycle from the instruction cache (I–Cache). It can dispatch up to six instructions per cycle through the dispatch unit, two of which are reserved for branch and compare instructions. Two indepen- dent branch processors within the ICU can each resolve one branch per cycle. Most of the branch operation penalties can be masked by resolving them concurrently with FXU and FPU instruction execution. The instruction cache is 32 Kbytes with 128 byte cache lines with a one cycle access time. The FXU decodes and executes all memory references, integer operations, and logical operations. It includes address trans- lation, data protection, and data cache directories for load/store instructions. There are two fixed–point execution units which are fed by their own instruction decode unit and maintain a copy of the general purpose register file. Two fixed–point instruc- tions can be executed per cycle by the two execution units. Each unit contains an adder and a logic functional unit providing addition, subtraction, and Boolean operations. One unit can also execute special operations such as cache operations and priv- ileged operations, while the other unit can perform fixed–point multiply and divide operations. The FPU consists of three units, a double precision arithmetic unit, a load unit, and a store unit. Each unit has dual pipelines and hence can execute up to two instructions per cycle. The peak issue rate from the FPU instruction buffers to these units is thus two double precision floating–point multiply–adds, two floating–point loads, and two floating–point stores per clock. The DCU includes a four–way set associative multiport write–back cache. The experiments in this paper were performed on a configuration with a 64 Kbytes data cache with 64 byte cache lines. The DCU also provides several buffers for cache and direct memory access (DMA) operations as well as error detection/correction and bit steering for all data sent to and received from memory. The DCU has a one cycle access time. The miss penalty to memory is determined experimentally to be between 16 and 21 processor clock cycles. 2.2. Interconnect Architecture [16][17] [18] The SP2 interconnect, termed the High–Performance Switch (HPS), is designed to minimize the average latency of message transmissions while allowing the aggregate bandwidth to scale linearly with the number of nodes. The HPS is a bidirectional 2 Instruction Cache Unit (ICU) Instruction Cache Dual Branch Dispatch Unit Processors Instruction buffers Instruction buffers Execution Execution Sync Unit without Unit with Arithmetic Store Load Multiply/ Multiply/ Execution Execution Execution Divide Divide Unit Unit Unit Fixed Point Unit (FXU) Floating Point Unit (FPU) Data Cache Unit (DCU) (4 Separate Chips) Secondary Cache Memory Unit (Optional) Figure 2: POWER2 Processor Architecture multistage interconnect (MIN), logically composed of switching elements.
Recommended publications
  • Evaluation of Architectural Support for Global Address-Based
    Evaluation of Architectural Supp ort for Global AddressBased Communication in LargeScale Parallel Machines y y Arvind Krishnamurthy Klaus E Schauser Chris J Scheiman Randolph Y Wang David E Culler and Katherine Yelick the sp ecic target architecture Wehave develop ed multi Abstract ple highly optimized versions of this compiler employing a Largescale parallel machines are incorp orating increas range of co degeneration strategies for machines with dedi ingly sophisticated architectural supp ort for userlevel mes cated network pro cessors In this studywe use this sp ec saging and global memory access We provide a systematic trum of runtime techniques to evaluate the p erformance evaluation of a broad sp ectrum of current design alternatives tradeos in architectural supp ort for communication found based on our implementations of a global address language in several of the current largescale parallel machines on the Thinking Machines CM Intel Paragon Meiko CS We consider ve imp ortant largescale parallel platforms Cray TD and Berkeley NOW This evaluation includes that havevarying degrees of architectural supp ort for com a range of compilation strategies that makevarying use of munication the Thinking Machines CM Intel Paragon the network pro cessor each is optimized for the target ar Meiko CS Cray TD and Berkeley NOW The CM pro chitecture and the particular strategyWe analyze a family vides direct userlevel access to the network the Paragon of interacting issues that determine the p erformance trade provides a network pro cessor
    [Show full text]
  • Parallel Computer Systems
    Parallel Computer Systems Randal E. Bryant CS 347 Lecture 27 April 29, 1997 Topics • Parallel Applications • Shared vs. Distributed Model • Concurrency Models • Single Bus Systems • Network-Based Systems • Lessons Learned Motivation Limits to Sequential Processing • Cannot push clock rates beyond technological limits • Instruction-level parallelism gets diminishing returns – 4-way superscalar machines only get average of 1.5 instructions / cycle – Branch prediction, speculative execution, etc. yield diminishing returns Applications have Insatiable Appetite for Computing • Modeling of physical systems • Virtual reality, real-time graphics, video • Database search, data mining Many Applications can Exploit Parallelism • Work on multiple parts of problem simultaneously • Synchronize to coordinate efforts • Communicate to share information – 2 – CS 347 S’97 Historical Perspective: The Graveyard • Lots of venture capital and DoD Resarch $$’s • Too big to enumerate, but some examples … ILLIAC IV • Early research machine with overambitious technology Thinking Machines • CM-2: 64K single-bit processors with single controller (SIMD) • CM-5: Tightly coupled network of SPARC processors Encore Computer • Shared memory machine using National microprocessors Kendall Square Research KSR-1 • Shared memory machine using proprietary processor NCUBE / Intel Hypercube / Intel Paragon • Connected network of small processors • Survive only in niche markets – 3 – CS 347 S’97 Historical Perspective: Successes Shared Memory Multiprocessors (SMP’s) • E.g., SGI
    [Show full text]
  • Scalability Study of KSR-1
    Scalability Study of the KSR-1 Appeared in Parallel Computing, Vol 22, 1996, 739-759 Umakishore Ramachandran Gautam Shah S. Ravikumar Jeyakumar Muthukumarasamy College of Computing Georgia Institute of Technology Atlanta, GA 30332 Phone: (404) 894-5136 e-mail: [email protected] Abstract Scalability of parallel architectures is an interesting area of current research. Shared memory parallel programming is attractive stemming from its relative ease in transitioning from sequential programming. However, there has been concern in the architectural community regarding the scalability of shared memory parallel architectures owing to the potential for large latencies for remote memory accesses. KSR-1 is a commercial shared memory parallel architecture, and the scalability of KSR-1 is the focus of this research. The study is conducted using a range of experiments spanning latency measurements, synchronization, and analysis of parallel algorithms for three computational kernels and an application. The key conclusions from this study are as follows: The communication network of KSR-1, a pipelined unidirectional ring, is fairly resilient in supporting simultaneous remote memory accesses from several processors. The multiple communication paths realized through this pipelining help in the ef®cient implementation of tournament-style barrier synchronization algorithms. Parallel algorithms that have fairly regular and contiguous data access patterns scale well on this architecture. The architectural features of KSR-1 such as the poststore and prefetch are useful for boosting the performance of parallel applications. The sizes of the caches available at each node may be too small for ef®ciently implementing large data structures. The network does saturate when there are simultaneous remote memory accesses from a fully populated (32 node) ring.
    [Show full text]
  • The KSR1: Experimentation and Modeling of Poststore Amy Apon Clemson University, [email protected]
    Clemson University TigerPrints Publications School of Computing 2-1993 The KSR1: Experimentation and Modeling of Poststore Amy Apon Clemson University, [email protected] E Rosti Universita degli studi de Milano E Smirni Vanderbilt University T D. Wagner Vanderbilt University M Madhukar Vanderbilt University See next page for additional authors Follow this and additional works at: https://tigerprints.clemson.edu/computing_pubs Part of the Computer Sciences Commons Recommended Citation Please use publisher's recommended citation. This Article is brought to you for free and open access by the School of Computing at TigerPrints. It has been accepted for inclusion in Publications by an authorized administrator of TigerPrints. For more information, please contact [email protected]. Authors Amy Apon, E Rosti, E Smirni, T D. Wagner, M Madhukar, and L W. Dowdy This article is available at TigerPrints: https://tigerprints.clemson.edu/computing_pubs/9 3 445b 0374303 7 E. Rasti E. Smirni A. W. Apoa L. w. Dowdy .- .. , . - . .. .. ... ..... i- ORNL/TM- 1228 7 I' Engineering Physics and Mathematics Division ; ?J -2 c_ Mathematical Sciences Section I.' THE KSR1: EXPERIMENTATION AND MODELING OF POSTSTORE E. Rosti E. Smirni t T. D. Wagner + A. W. Apon L. W. Dowdy Dipartimento di Scienze dell'Informazione Universitb degli Studi di Milano Via Comelico 39 20135 Milano, Italy t Computer Science Department Vaiiderbilt University Box 1679, Station B Nashville, TN 37235 Date Published: February 1993 This work was partially supported by sub-contract 19X-SL131V from the Oak Ridge National Laboratory, and by grant N. 92.01615.PF69 from the Italian CNR "Progetto Finalizzato Sistemi Informatici e Calcolo Parallel0 - Sottoprogetto 3." Prepared by the Oak Ridge National Laboratory Oak Ridge, Tennessee 37831 managed by Martin Marietta Energy Systems, Inc.
    [Show full text]
  • THE RISE and Fall the 01 BRILLIANT START-UP THAT Some Day We Will Build a Think­ I~Z~~~~~ Thinking Ing Machine
    Company Profile THE RISE and Fall THE 01 BRILLIANT START-UP THAT Some day we will build a think­ I~Z~~~~~ Thinking ing machine. It will be a truly NEVER GRASPED intelligent machine. One that can see and hear and speak. A THE BASICS Mach-Ines machine that will be proud of us. by Gary Taubes -From a Thinking Machines brochure seven 'years a~ter. its The truth is very different. This is the simple proeessors, all of them completing In 19 90 founding, Thlllklllg story of how Thinking Machines got the a single instruction at the same time. To Machines was the market leader in paral­ jump on a hot new market-and then get more speed, more processors would lel supercomputers, with sales of about screwed up, big time. be added. Eventually, so the theory went, $65 million. Not only was the company with enough processors (perhaps billions) protitable; it also, in the words of one IBM ntil W. Daniel Hillis came along, and the right software, a massively paral­ computer scientist, had cornered the mar­ Ucomputers more or less had been de­ lel computer might start acting vaguely . ket "on sex appeal in high-performance signed along the lines of ENIAC. Ifl that human. Whether it would take pride in its computing." Several giants in the com­ machine a single processor complete? in­ creators would remain to be seen. puter industry were seeking a merger or a structions one at a time, in sequence. "Se­ Hillis is what good scientists call a very partnership with the company.
    [Show full text]
  • NASA Contractor Report ICASE Report No. 94-2 191592 SHARED
    //v NASA Contractor Report 191592 ICASE Report No. 94-2 IC S SHARED VIRTUAL MEMORY AND GENERALIZED SPEEDUP (NASA-CR-1915q2) SHARED VIRTUAL N94-27416 MEMORY AND GENERALIZED SPEEDUP Final Report (ICASE) 24 p Xian-He Sun Unclas Jianping Zhu G3/61 0000316 NASA Contract No. NAS 1-19480 January 1994 Institute for Computer Applications in Science and Engineering NASA Langley Research Center Hampton, Virginia 23681-0001 Operated by the Universities Space Research Association National Aeronautics and Space Administration Langley Research Center Hampton, Virginia 23681-0001 Shared Virtual Memory and Generalized Speedup * Xian-He Sun Jianping Zhu ICASE NSF Engineering Research Center Mail Stop 132C Deft. of Math. and Stat. NASA Langley Research Center Mississippi State University Hampton, VA 23681-0001 Mississippi State, MS 39762 Abstract Generalized speedup is defined as parallel speed over sequential speed. In this pa- per the generalized speedup and its relation with other existing performance metrics, such as traditional speedup, efficiency, scalability, etc., are carefully studied. In terms of the introduced asymptotic speed, we show that the difference between the gener- alized speedup and the traditional speedup lies in the definition of the efficiency of uniprocessor processing, which is a very important issue in shared virtual memory ma- chines. A scientific application has been implemented on a KSR-1 parallel computer. Experimental and theoretical results show that the generalized speedup is distinct from the traditional speedup and provides a more reasonable measurement. In the study of different speedups, various causes of superlinear speedup are also presented. "This research was supported by the National Aeronautics and Space Administration under NASA contract NAS1- 19480 while the first author was in residence at the Institute for Computer Applications in Science and Engineering (ICASE), NASA Langley Research Center, Hampton, VA 23681-0001.
    [Show full text]
  • NSA and the Supercomputer Industry
    ·DOCID: 4001121 I OP S!Cltl!!T t:IMIAll: . NSA and the Supercomputer Industry · (b)(1 ) -(Eij"(3J::PT:-00°36--------------- -- ------- -----------------------------_I______ _ (b)(3)-50 USC 403 (b)(3)-P.L. 86-36 NSA'S RELIANCE ON BIGH-PERFORMANC,ECOMPUTING TECHNOLOGY l'rsc.i.. NSA is heavily depende.nt -~p~n high-performance computing m·J>C) technology. heretofore known as ~pe~mputing technology, particularly in the area\of cryptanalysis. HPC is used for' fundamental cryptanalytic CCA) mathematical research. diagnosis of unknown ~tofogics, development of attacks, and daily exploitation fot. OO~n. fut~ u~:·:-m1::e SIGINT production_.. CA efforts which eventually tbeir way p:tn \. devices (SPl),s);·high-performance desktop computers, ort _ _ ___ __ I _______fiad their genesis in breakthroughs made on NSA: supercompu rs. tiona y, NSA has a smaller portion of its HPC assets devoted to signals processing and to the protection of United States cryptographic aystems. ~ As a consequence, NSA has constructed the largest single-site, single­ mission supercomputer complex in the world. This has allowed the Agency to gain unique HPC experience as well as leverage with HPC vendors. A symbiotic relationship has evolved over time. NSA shares probleme and requirements with HPC vendors~ in exchange it gets systems that not only meet its needs better but also create improved products for the entire range of HPC cust.omers. Indeed, NSA can c;ite multiple examples of supercomputer technology spanning several decades which was designed as a direct result ofour request&. ~Of equal importance is technology related to the construction of SPDs.
    [Show full text]
  • Appendix M Historical Perspectives and References
    M.1 Introduction M-2 M.2 The Early Development of Computers (Chapter 1) M-2 M.3 The Development of Memory Hierarchy and Protection (Chapter 2 and Appendix B) M-9 M.4 The Evolution of Instruction Sets (Appendices A, J, and K) M-17 M.5 The Development of Pipelining and Instruction-Level Parallelism (Chapter 3 and Appendices C and H) M-27 M.6 The Development of SIMD Supercomputers, Vector Computers, Multimedia SIMD Instruction Extensions, and Graphical Processor Units (Chapter 4) M-45 M.7 The History of Multiprocessors and Parallel Processing (Chapter 5 and Appendices F, G, and I) M-55 M.8 The Development of Clusters (Chapter 6) M-74 M.9 Historical Perspectives and References M-79 M.10 The History of Magnetic Storage, RAID, and I/O Buses (Appendix D) M-84 M Historical Perspectives and References If … history … teaches us anything, it is that man in his quest for knowledge and progress is determined and cannot be deterred. John F. Kennedy Address at Rice University (1962) Those who cannot remember the past are condemned to repeat it. George Santayana The Life of Reason (1905), Vol. 2, Chapter 3 M-2 ■ Appendix M Historical Perspectives and References M.1 Introduction This appendix provides historical background on some of the key ideas presented in the chapters. We may trace the development of an idea through a series of machines or describe significant projects. If you are interested in examining the initial development of an idea or machine or are interested in further reading, references are provided at the end of each section.
    [Show full text]
  • National Energy Research Scientific Computing Center (NERSC) The
    National Energy Research Scientific Computing Center (NERSC) The Divergence Problem Horst D. Simon Director, NERSC Center Division, LBNL November 19, 2002 # Outline ? Introducing NERSC-3 E ? The Divergence Problem ? What NERSC is doing about it Combined NERSC-3 Characteristics ? The combined NERSC-3/4 system (NERSC-3Base and NERSC-3Enhanced) will have — 416 16 way Power 3+ nodes with each CPU at 1.5 Gflop/s ?380 for computation — 6,656 CPUs – 6,080 for computation — Total Peak Performance of 10 Teraflop/s — Total Aggregate Memory is 7.8 TB — Total GPFS disk will be 44 TB ?Local system disk is an additional 15 TB — Combined SSP-2 measure is 1.238 Tflop/s — NERSC-3E be in production by the end of Q1/CY03 ?Nodes will arrive in the first two weeks of November Comparison with Other Systems NERSC-3 E ASCI White ES Cheetah PNNL (ORNL) Mid 2003 Nodes 416 512 640 27 700 CPUs 6,656 8,192 5,120 864 1400 Peak(Tflops) 10 12 40 4.5 9.6(8.3) Memory (TB) 7.8 4 10 1 1.8 Disk(TB) 60 150 700 9 53 SSP(Gflop/s) 1,238 1,652 179 PNNL system available in Q3 CY2003 SSP = sustained systems performance (NERSC applications benchmark) Outline ? Introducing NERSC-3 E ? The Divergence Problem ? What NERSC is doing about it Signposts of Change in HPC In early 2002 there were several signposts, which signal a fundamental change in HPC in the US: ? Installation and very impressive early performance results of the Earth Simulator System (April 2002) ? Lack of progress in computer architecture research evident at Petaflops Workshop (WIMPS, Feb.
    [Show full text]
  • Hardware for Fast Global Operations on Distributed Memory Multicomputers and Multiprocessors
    Portland State University PDXScholar Dissertations and Theses Dissertations and Theses 1995 Hardware for Fast Global Operations on Distributed Memory Multicomputers and Multiprocessors Douglas V. Hall Portland State University Follow this and additional works at: https://pdxscholar.library.pdx.edu/open_access_etds Part of the Electrical and Computer Engineering Commons Let us know how access to this document benefits ou.y Recommended Citation Hall, Douglas V., "Hardware for Fast Global Operations on Distributed Memory Multicomputers and Multiprocessors" (1995). Dissertations and Theses. Paper 1286. https://doi.org/10.15760/etd.1285 This Dissertation is brought to you for free and open access. It has been accepted for inclusion in Dissertations and Theses by an authorized administrator of PDXScholar. Please contact us if we can make this document more accessible: [email protected]. HARDWARE FOR FAST GLOBAL OPERATIONS ON DISTRIBUTED MEMORY MULTICOMPUTERS AND MULTIPROCESSORS by DOUGLAS VINCENT HALL A submitted in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY In ELECTRICAL AND COMPUTER ENGINEERING Portland State University 1995 DISSERTATION APPROVAL The abstract and dissertation of Douglas Vincent Hall for the Doctor of Philosophy in Electrical and Computer Engineering were presented December 9, 1994, and accepted by dissertation committee and the doctora~,rogra. m.' /; COMMITTEE APPROVALS: ~------ Michael A. Driscoll, ~ , Marek Perkowski / W, Robert Daasch Richard D, Morris Representative of the Office of Graduate Studies DOCTORAL PROGRAM APPROVAL: Rolf Schaumann, Chair Department of Electrical Engineering ************************************************************************ ACCEPTED FOR PORTLAND STATE UNIVERSITY BY THE LIBRARY by ABSTRACT An abstract of the dissertation of Douglas Vincent Hall for the Doctor of Philosophy in Electrical and Computer Engineering presented December 9, 1994.
    [Show full text]
  • A Survey of Parallel Programming Languages and Tools
    ASurveyofParallel Programming Languages and Tools Doreen Y.Cheng Report RND-93-005 March 1993 NASA Ames Research Center M/S 256-6 Moffett Field, CA 94035 -2- Abstract This survey examines thirty-fiveparallel programming languages and fifty-nine par- allel programming tools. It focuses on tool capabilities needed for writing parallel scien- tific programs, not on features that explore general computer science issues. The tools are classified based on their functions and ranked with current and future needs of NAS in mind: in particular,existing and anticipated NAS supercomputers and workstations, oper- ating systems, programming languages, and applications. The report is designed to give readers a quick grasp of the tool features, and provides tables to compare their main func- tions. -3- Introduction Providing sufficient parallel programming tools is one key step to enable NAS users to use parallel computers. The report "A Survey ofParallel Programming Tools" (RND-91-005) has been requested by more than 100 organizations. Since then, NAS has added massively parallel supercomputers into its production computing facilities. Pro- gramming for a wide variety of parallel architectures (SIMD, shared-memory MIMD, distributed-memory MIMD, heterogeneous MIMD) demands a survey ofabroader range of programming tools than the tools included in report RND-91-005. In response to the newdemand, this report surveysparallel programming languages as well as tools. In the text below, theyare both referred to as tools. The scope of the survey has been enlarged to include tools for all forms of parallel architectures and programming paradigms. More than 80 tools were submitted to the survey;only a fewwere eliminated due to their proprietary or obsolescent nature.
    [Show full text]
  • A History of Modern 64-Bit Computing
    A History of Modern 64-bit Computing Matthew Kerner [email protected] Neil Padgett [email protected] CSEP590A Feb 2007 -1- Background By early 2002, the rumors had been swirling for months: despite the company’s intense focus on its IA64 64-bit processor, Intel engineers in Oregon were busy updating the company’s venerable x86/Pentium line with 64-bit EM64T (later Intel64) extensions. Not only would this mean Intel was potentially forgoing, or at least hedging, its bet on IA64 as its next-generation processor and entrance into the 64-bit processor space; reports *1+ were also claiming that Intel’s 64-bit x86 processor would be compatible with the x86-64 (later AMD64) architecture that rival AMD had announced just over a year before [2]. Intel boasted revenues of $26.5 Billion in 2001 [3], and was a giant compared to AMD, who sold a relatively tiny $3.89 Billion that year [4]. If true, Intel would be making a bold admission of AMD’s coup by following the technology lead of its smaller rival. This run-in was hardly the first time Intel and AMD crossed paths however; both had deep roots in Silicon Valley and in some ways they shared a common history. The relationship was a tortured one: the two companies were alternately friends advancing the x86 architecture and foes competing furiously. Intel’s founding traces back through Fairchild Semiconductor to Shockley Semiconductor Lab, the post-Bell Labs home of transistor co-inventor William Shockley. Founded in 1955 [5], Shockley Labs focused on the development and evolution of the transistor.
    [Show full text]